傳送端於非完美通道資訊下之細胞間干擾處理

Abstract

在未來的無線通訊中，越來越多的小型基地台 (small cell) 將會被布建，故其互相干擾的機會將會提高。因此細胞間干擾 (inter-cell interference) 將會是未來無線通訊中的一大議題。從消息理論 (information theory) 的觀點，最簡單的互干擾連結是一個兩用戶的干擾通道模型 (2-user interference channel) 。在此模型下，目前Han-Kobayashi部分干擾消除法 (partial interference cancellation) 可以達到最好的傳輸率區域。而其需要完美 (精準且全盤) 的通道資訊 (perfect channel state information) 才能予以實踐，然而通道資訊通常會因通道估計誤差 (estimation error) 及所在位置 (locality) 導致其是不精準且局部性的。因此上述部分干擾消除法的效能將會無可避免的降低。 在此篇論文中，我們分析了在兩用戶干擾通道下，且有非完美通道資訊時，使用上述部分干擾消除法的系統效能。除此之外，我們利用局部性的通道資訊，於衰落 (fading) 干擾通道下提出一結合了部分干擾消除法優點的動態功率分配演算法 (dynamic power allocation algorithm)。並藉由模擬，展示出其跟傳統干擾處理方法相比，效能將會大幅的提升。我們並將此概念延伸到有細胞間干擾的上行通道行動通訊系統(uplink cellular system)。並搭配連續干擾消除 (successive interference cancellation) 及使用者排程 (user-scheduling)，分析其效能。

For future wireless communication, more and more small cells will be established, there are more likely to interfere each other. Thus inter-cell interference management will be an important issue. From the point of view of information theory, the simplest interfering links can be modeled as a two-user interference channel (IC). For this IC model, a partial interference cancellation (PIC) method called Han-Kobayashi scheme can achieve the best-known achievable rate region. However, it needs perfect (precise and global) channel state information (CSI). For practical communication, CSI is usually imperfect (imprecise and local) due to channel estimation error and locality respectively. Hence the performance of the PIC scheme will be reduced inevitably. For this thesis, we analyze the performance of two-user IC with the PIC scheme and imprecise CSI. In addition, we utilize the local CSI to propose a dynamic power allocation algorithm combined the PIC scheme for fading IC. It's shown that the performance is obviously enhanced compared with some traditional schemes. We further extend above result to uplink cellular with inter-cell interference, and combine the successive interference cancellation (SIC) and user-scheduling to analyze its performance.